Skip to contentSkip to conversation

fs/writeback: skip AS_NO_DATA_INTEGRITY mappings in wait_sb_inodes()

PATCH Applied Joanne Koong 2 participants 3 messages Jan 5 – Jan 6

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

v3 cover fs/writeback: skip AS_NO_DATA_INTEGRITY mappings in wait_sb_inodes()

This patch adds a new mapping flag AS_NO_DATA_INTEGRITY which signifies that a mapping does not have data integrity guarantees, and skips waiting on writeout for these mappings in wait_sb_inodes(), as these mappings cannot guarantee that data is persistently stored. This patch sets this flag on fuse mappings.

This fixes the userspace regression reported by Athul and J. upstream in [1][2] where if there is a bug in a fuse server that causes the server to never complete writeback, it will make wait_sb_inodes() wait forever.

Thanks,
Joanne

[1] https://lore.kernel.org/regressions/CAJnrk1ZjQ8W8NzojsvJPRXiv9TuYPNdj8Ye7=Cgkj=iV_i8EaA@mail.gmail.com/T/#t
[2] https://lore.kernel.org/linux-fsdevel/[email protected]/

Changelog:
v2: https://lore.kernel.org/linux-fsdevel/[email protected]/
* Add comments to commit message (David) and to wait_sb_inodes() (Andrew)
* Add Bernd's Reviewed-by and J's Tested-by

v1: https://lore.kernel.org/linux-mm/[email protected]/
* Change AS_WRITEBACK_MAY_HANG to AS_NO_DATA_INTEGRITY and keep AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM as is.

Joanne Koong (1):
fs/writeback: skip AS_NO_DATA_INTEGRITY mappings in wait_sb_inodes()

 fs/fs-writeback.c       |  7 ++++++-
 fs/fuse/file.c          |  4 +++-
 include/linux/pagemap.h | 11 +++++++++++
 3 files changed, 20 insertions(+), 2 deletions(-)
Signature · 1 line
2.47.3
#1 lore raw

PATCH v3 1/1 fs/writeback: skip AS_NO_DATA_INTEGRITY mappings in wait_sb_inodes()

Above the while() loop in wait_sb_inodes(), we document that we must wait for all pages under writeback for data integrity. Consequently, if a mapping, like fuse, traditionally does not have data integrity semantics, there is no need to wait at all; we can simply skip these inodes.

This restores fuse back to prior behavior where syncs are no-ops. This fixes a user regression where if a system is running a faulty fuse server that does not reply to issued write requests, this causes wait_sb_inodes() to wait forever.

Fixes 0c58a97f919c ("fuse: remove tmp folio for writebacks and internal rb tree")
---
 fs/fs-writeback.c       |  7 ++++++-
 fs/fuse/file.c          |  4 +++-
 include/linux/pagemap.h | 11 +++++++++++
 3 files changed, 20 insertions(+), 2 deletions(-)
fs/fs-writeback.c +6 -1
index 6800886c4d10..baa2f2141146 100644
@@ -2750,8 +2750,13 @@ static void wait_sb_inodes(struct super_block *sb)
 		 * The mapping can appear untagged while still on-list since we
 		 * do not have the mapping lock. Skip it here, wb completion
 		 * will remove it.
+		 *
+		 * If the mapping does not have data integrity semantics,
+		 * there's no need to wait for the writeout to complete, as the
+		 * mapping cannot guarantee that data is persistently stored.
 		 */
-		if (!mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK))
+		if (!mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK) ||
+		    mapping_no_data_integrity(mapping))
 			continue;
 
 		spin_unlock_irq(&sb->s_inode_wblist_lock);
fs/fuse/file.c +3 -1
index 01bc894e9c2b..3b2a171e652f 100644
@@ -3200,8 +3200,10 @@ void fuse_init_file_inode(struct inode *inode, unsigned int flags)
 
 	inode->i_fop = &fuse_file_operations;
 	inode->i_data.a_ops = &fuse_file_aops;
-	if (fc->writeback_cache)
+	if (fc->writeback_cache) {
 		mapping_set_writeback_may_deadlock_on_reclaim(&inode->i_data);
+		mapping_set_no_data_integrity(&inode->i_data);
+	}
 
 	INIT_LIST_HEAD(&fi->write_files);
 	INIT_LIST_HEAD(&fi->queued_writes);
include/linux/pagemap.h +11 -0
index 31a848485ad9..ec442af3f886 100644
@@ -210,6 +210,7 @@ enum mapping_flags {
 	AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM = 9,
 	AS_KERNEL_FILE = 10,	/* mapping for a fake kernel file that shouldn't
 				   account usage to user cgroups */
+	AS_NO_DATA_INTEGRITY = 11, /* no data integrity guarantees */
 	/* Bits 16-25 are used for FOLIO_ORDER */
 	AS_FOLIO_ORDER_BITS = 5,
 	AS_FOLIO_ORDER_MIN = 16,
@@ -345,6 +346,16 @@ static inline bool mapping_writeback_may_deadlock_on_reclaim(const struct addres
 	return test_bit(AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM, &mapping->flags);
 }
 
+static inline void mapping_set_no_data_integrity(struct address_space *mapping)
+{
+	set_bit(AS_NO_DATA_INTEGRITY, &mapping->flags);
+}
+
+static inline bool mapping_no_data_integrity(const struct address_space *mapping)
+{
+	return test_bit(AS_NO_DATA_INTEGRITY, &mapping->flags);
+}
+
 static inline gfp_t mapping_gfp_mask(const struct address_space *mapping)
 {
 	return mapping->gfp_mask;
Signature · 1 line
2.47.3
maintainer
#2 lore raw
Quoted 11 lines · On Mon, 5 Jan 2026 13:17:27 -0800 Joanne Koong <[email protected]> wrote:

Above the while() loop in wait_sb_inodes(), we document that we must wait for all pages under writeback for data integrity. Consequently, if a mapping, like fuse, traditionally does not have data integrity semantics, there is no need to wait at all; we can simply skip these inodes.

This restores fuse back to prior behavior where syncs are no-ops. This fixes a user regression where if a system is running a faulty fuse server that does not reply to issued write requests, this causes wait_sb_inodes() to wait forever.

Quoted 7 lines
Fixes 0c58a97f919c ("fuse: remove tmp folio for writebacks and internal rb tree")

Thanks.

Quoted 4 lines
 fs/fs-writeback.c       |  7 ++++++-
 fs/fuse/file.c          |  4 +++-
 include/linux/pagemap.h | 11 +++++++++++
 3 files changed, 20 insertions(+), 2 deletions(-)

I'll queue this in mm.git's mm-hotfixes branches for a 6.19-rcX merge.